U.S. patent application number 14/514090 was filed with the patent office on 2015-01-29 for photoelectric conversion device, image pickup system and method of manufacturing photoelectric conversion device.
The applicant listed for this patent is CANON KABUSHIKI KAISHA. Invention is credited to Takeshi Ichikawa, Yasuhiro Sekine, Mineo Shimotsusa.
Application Number | 20150031162 14/514090 |
Document ID | / |
Family ID | 49001931 |
Filed Date | 2015-01-29 |
United States Patent
Application |
20150031162 |
Kind Code |
A1 |
Shimotsusa; Mineo ; et
al. |
January 29, 2015 |
PHOTOELECTRIC CONVERSION DEVICE, IMAGE PICKUP SYSTEM AND METHOD OF
MANUFACTURING PHOTOELECTRIC CONVERSION DEVICE
Abstract
A photoelectric conversion device includes a first semiconductor
substrate including a photoelectric conversion unit for generating
a signal charge in accordance with an incident light, and a second
semiconductor substrate including a signal processing unit for
processing an electrical signal on the basis of the signal charge
generated in the photoelectric conversion unit. The signal
processing unit is situated in an orthogonal projection area from
the photoelectric conversion unit to the second semiconductor
substrate. A multilayer film including a plurality of insulator
layers is provided between the first semiconductor substrate and
the second semiconductor substrate. The thickness of the second
semiconductor substrate is smaller than 500 micrometers. The
thickness of the second semiconductor substrate is greater than the
distance from the second semiconductor substrate and a
light-receiving surface of the first semiconductor substrate.
Inventors: |
Shimotsusa; Mineo;
(Machida-shi, JP) ; Ichikawa; Takeshi;
(Hachioji-shi, JP) ; Sekine; Yasuhiro;
(Yokohama-shi, JP) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
CANON KABUSHIKI KAISHA |
Tokyo |
|
JP |
|
|
Family ID: |
49001931 |
Appl. No.: |
14/514090 |
Filed: |
October 14, 2014 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
13774150 |
Feb 22, 2013 |
8890331 |
|
|
14514090 |
|
|
|
|
Current U.S.
Class: |
438/67 |
Current CPC
Class: |
H01L 27/14687 20130101;
H01L 27/14621 20130101; H01L 27/14634 20130101; H01L 31/024
20130101; H01L 2224/02166 20130101; H01L 2224/04042 20130101; H01L
27/1469 20130101; H01L 2924/351 20130101; H01L 2924/12043 20130101;
H01L 2924/13091 20130101; H01L 27/14618 20130101; H01L 2224/48463
20130101; H01L 2924/12042 20130101; H01L 24/05 20130101; H01L
2924/13091 20130101; H01L 2924/00 20130101; H01L 2224/04042
20130101; H01L 2924/00 20130101; H01L 2924/351 20130101; H01L
2924/00 20130101; H01L 2924/12042 20130101; H01L 2924/00 20130101;
H01L 2924/12043 20130101; H01L 2924/00 20130101 |
Class at
Publication: |
438/67 |
International
Class: |
H01L 27/146 20060101
H01L027/146 |
Foreign Application Data
Date |
Code |
Application Number |
Feb 29, 2012 |
JP |
2012-043962 |
Claims
1. A method of manufacturing a photoelectric conversion device,
comprising: a bonding process in which a first member and a second
member are bonded to each other to prepare a composite member, the
first member including a first semiconductor wafer on which a
plurality of photoelectric conversion element groups are arranged
and a first film including an insulator layer and provided on a
front surface of the first semiconductor wafer, and the second
member including a second semiconductor wafer on which a plurality
of semiconductor device groups are arranged and a second film
including an insulator layer and provided on a front surface of the
second semiconductor wafer, the first member and the second member
are bonded with the first film and the second film being disposed
between the first semiconductor wafer and the second semiconductor
wafer: a first thinning process in which, after the bonding
process, the first semiconductor wafer of the composite member is
thinned from a back surface side of the first semiconductor wafer;
and a second thinning process in which, after the first thinning
process, the second semiconductor wafer of the composite member is
thinned from a back surface side of the second semiconductor wafer,
wherein the thickness of the second semiconductor wafer after the
second thinning process is greater than the distance between the
back surface of the first semiconductor wafer after the first
thinning process and the front surface of the second semiconductor
wafer.
2. The method of manufacturing a photoelectric conversion device
according to claim 1, further comprising, after the second thinning
process, a dividing process in which the composite member is
divided into a plurality of photoelectric conversion devices each
including the photoelectric conversion element group and the
semiconductor device group.
3. The method of manufacturing a photoelectric conversion device
according to claim 1, wherein the thickness of the second
semiconductor wafer after the second thinning process is equal to
or greater than 10 times the thickness of the first semiconductor
wafer after the first thinning process and is equal to or smaller
than 100 times the thickness of the first semiconductor wafer after
the first thinning process.
4. The method of manufacturing a photoelectric conversion device
according to claim 1, wherein: in the bonding process, the
thickness of the first semiconductor wafer and the second
semiconductor wafer are equal to or greater than 500 micrometers;
in the first thinning process, the thickness of the first
semiconductor wafer is set to equal to or smaller than 10
micrometers; and in the second thinning process, the thickness of
the second semiconductor wafer is set to equal to or smaller than
400 micrometers.
5. The method of manufacturing a photoelectric conversion device
according to claim 2, further comprising: between the first
thinning process and the second thinning process, a process to form
a third film on the first semiconductor wafer; and after the
dividing process, a process to fix a surface of the semiconductor
device on the side opposite to the third film to a heat-conductive
member having thermal conductivity which is higher than thermal
conductivity of the second semiconductor wafer.
6. The method of manufacturing a photoelectric conversion device
according to claim 2, further comprising, after the second thinning
process and before the dividing process, a process to form a recess
on a back surface of the second semiconductor substrate.
7. The method of manufacturing a photoelectric conversion device
according to claim 2, wherein the dividing process is performed by
a laser dicing.
8. The method of manufacturing a photoelectric conversion device
according to claim 1, wherein the thickness of the second
semiconductor wafer after the second thinning process is set to
equal to or greater than 20 micrometers and is set to equal to or
smaller than 500 micrometers.
9. The method of manufacturing a photoelectric conversion device
according to claim 1, wherein the bonding process is performed so
that a signal processing unit constituted by the one of the
plurality of semiconductor device groups is situated in an
orthogonal projection area from one of the plurality of
photoelectric conversion element groups to the second semiconductor
wafer, wherein the thickness of the second semiconductor wafer
after the second thinning process is set to equal to or greater
than 50 micrometers.
10. The method of manufacturing a photoelectric conversion device
according to claim 1, wherein the second thinning process is
performed by at least a mechanical polishing.
11. The method of manufacturing a photoelectric conversion device
according to claim 1, wherein the second thinning process is
performed by a mechanical polishing and a chemical mechanical
polishing, and the thickness of the second semiconductor wafer
after the second thinning process is set to equal to or smaller
than 300 micrometers.
Description
CROSS-REFERENCE TO RELATED APPLICATION
[0001] This application is a Divisional of U.S. application Ser.
No. 13/774,150, filed Feb. 22, 2013, which claims priority from
Japanese Patent Application No. 2012-043962 filed Feb. 29, 2012,
which are hereby incorporated by reference herein in their
entireties.
BACKGROUND OF THE INVENTION
[0002] 1. Field of the Invention
[0003] One disclosed aspect of the embodiments relates to a
photoelectric conversion device which includes a plurality of
semiconductor substrates.
[0004] 2. Description of the Related Art
[0005] In a related art photoelectric conversion device, a
photoelectric conversion unit including a plurality of
photoelectric conversion elements, a signal processing unit for
processing an electrical signal from the photoelectric conversion
unit, and a control unit for controlling the photoelectric
conversion unit and the signal processing unit are monolithically
mounted on a single semiconductor substrate. It has been demanded
to increase an area of the photoelectric conversion unit, to reduce
the area of the photoelectric conversion device and to achieve
higher signal processing capability. In Japanese Patent Laid-Open
No. 2011-159958, it is discussed to stack a plurality of
semiconductor substrates, to provide a photoelectric conversion
unit in one of the semiconductor substrates and provide a signal
processing unit in the other of the semiconductor substrate.
[0006] Since the signal processing unit is driven at higher speed
(at higher frequency) than the photoelectric conversion unit, the
power consumed in the signal processing unit is larger than that
consumed in the photoelectric conversion unit. Generated heat
becomes greater in proportion to the consumed power. If a plurality
of semiconductor substrates are stacked, since these semiconductor
substrates are positioned close to each other, an influence of
generation of heat of the signal processing unit becomes
significant as compared with a case in which the photoelectric
conversion unit and the signal processing unit are mounted
monolithically on a single semiconductor substrate.
SUMMARY OF THE INVENTION
[0007] A first aspect of the technique for solving the
aforementioned problem is a photoelectric conversion device which
includes: a first semiconductor substrate which includes a
photoelectric conversion unit for generating signal charge in
accordance with incident light; and a second semiconductor
substrate which includes a signal processing unit for processing an
electrical signal on the basis of the signal charge, the signal
processing unit being situated in an orthogonal projection area
from the photoelectric conversion unit to the second semiconductor
substrate, and a multilayer film including a plurality of insulator
layers being provided between the first semiconductor substrate and
the second semiconductor substrate, wherein the thickness of the
second semiconductor substrate is smaller than 500 micrometers, and
the thickness of the second semiconductor substrate is greater than
the distance between the second semiconductor substrate and a
light-receiving surface of the first semiconductor substrate.
[0008] A second aspect of the technique for solving the
aforementioned problem is a method of manufacturing a photoelectric
conversion device which includes: a bonding process in which a
first member including a first semiconductor wafer on which a
plurality of photoelectric conversion element groups are arranged
and a first film including an insulator layer and provided on a
front surface of the first semiconductor wafer, and a second member
including a second semiconductor wafer on which a plurality of
semiconductor device groups are arranged and a second film
including an insulator layer and provided on a front surface of the
second semiconductor wafer are bonded to each other with the first
film and the second film being disposed between the first
semiconductor wafer and the second semiconductor wafer to prepare a
composite member: a first thinning process in which, after the
bonding process, the first semiconductor wafer of the composite
member is thinned from a back surface side of the first
semiconductor wafer; and a second thinning process in which, after
the first thinning process, the second semiconductor wafer of the
composite member is thinned from a back surface side of the second
semiconductor wafer, wherein: the thickness of the second
semiconductor wafer after the thinning is smaller than 500
micrometers and the thickness of the second semiconductor wafer
after the thinning is greater than the distance between the back
surface of the first semiconductor wafer after the thinning and the
surface of the second semiconductor wafer.
[0009] According to those techniques, a photoelectric conversion
device in which the influence of generation of heat of the signal
processing unit is reduced may be provided.
[0010] Further features of the disclosure will become apparent from
the following description of exemplary embodiments with reference
to the attached drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] FIGS. 1A to 1C are schematic diagrams of a main part of a
photoelectric conversion device.
[0012] FIG. 2 is a schematic diagram of a main part of
photoelectric conversion device.
[0013] FIGS. 3A-1 to 3D-2 are schematic diagrams of the
photoelectric conversion device.
[0014] FIGS. 4A and 4B are schematic diagrams of the photoelectric
conversion device.
[0015] FIGS. 5A to 5J are schematic diagrams of a method of
manufacturing the photoelectric conversion device.
DESCRIPTION OF THE EMBODIMENTS
[0016] Hereinafter, an exemplary embodiment will be described with
reference to the drawings. In the following description and
drawings, a plurality of drawings may be mutually referred to. The
same or similar configurations are denoted by common reference
numerals and description of configurations denoted by the common
reference numerals is omitted appropriately.
First Embodiment
[0017] A photoelectric conversion device will be described briefly.
FIG. 1A is a perspective view of a semiconductor device 1 which is
a main part of the photoelectric conversion device. FIGS. 1B and 1C
are exploded perspective views of an exemplary semiconductor device
1. As illustrated in FIG. 1A or 1B, in the semiconductor device 1
illustrated in FIG. 1A, a first semiconductor substrate 10 is
stacked on a second semiconductor substrate 20 in an
electrically-connected manner.
[0018] The semiconductor device 1 includes a first semiconductor
substrate 10 which includes a photoelectric conversion unit 11 in
which signal charge is generated in accordance with incident light.
The photoelectric conversion unit 11 is constituted by a plurality
of photoelectric conversion elements which are arranged in two
dimensions. The photoelectric conversion elements may be
photodiodes and photogates. The photoelectric conversion unit 11
may include a signal generating circuit which generates an
electrical signal on the basis of the signal charge generated in
the photoelectric conversion element. The signal generating circuit
may be constituted by a transfer transistor, an amplifying
transistor and a reset transistor.
[0019] The semiconductor device 1 includes a second semiconductor
substrate 20 which includes a signal processing unit 22. The signal
processing unit 22 processes the electrical signal on the basis of
the signal charge generated in the photoelectric conversion unit
11. In FIG. 1A, the photoelectric conversion unit 11 is surrounded
by a dash-dot line and the signal processing unit 22 is surrounded
by a dash-dot-dot line. The signal processing unit 22 is situated
in an orthogonal projection area from the photoelectric conversion
unit 11 to the second semiconductor substrate 20. The signal
processing unit 22 may include a noise reduction circuit, an
amplifier circuit, a conversion circuit and an image signal
processing circuit. The noise reduction circuit is, for example, a
correlation double sampling (CDS) circuit. The amplifier circuit
is, for example, a sequential amplifier circuit. The conversion
circuit is, for example, an analog-to-digital conversion (ADC)
circuit constituted by a comparator and a counter. The image signal
processing circuit, which includes, for example, memory and a
processor, generates image data from digital signals which has been
converted from analog signals, and performs image processing to the
image data. The entire or a part of the signal processing unit 22
may be situated in the entire orthogonal projection area from the
photoelectric conversion unit 11 to the second semiconductor
substrate 20. Alternatively, the entire or a part of the signal
processing unit 22 may be situated in a part of the orthogonal
projection area from the photoelectric conversion unit 11 to the
second semiconductor substrate 20. FIG. 1A illustrates an example
in which a part of the signal processing unit 22 is situated in a
part of the orthogonal projection area from the photoelectric
conversion unit 11 to the second semiconductor substrate 20. A part
of the signal processing unit 22 may be provided in the first
semiconductor substrate 10. For example, the signal processing unit
for analog signals, such as the noise reduction circuit and the
amplifier circuit, may be provided in the first semiconductor
substrate 10, and the signal processing unit for digital signals,
such as the conversion circuit and the image signal processing
circuit, may also be provided in the second semiconductor substrate
20.
[0020] A multilayer film 30 including a plurality of insulator
layers is provided between the first semiconductor substrate 10 and
the second semiconductor substrate 20. The multilayer film 30 may
include a plurality of conductor layers. The conductor layer of the
multilayer film 30 functions as wiring of an electric circuit
provided in the first semiconductor substrate 10 and the second
semiconductor substrate 20, and wiring which connects an electric
circuit provided in the first semiconductor substrate 10 and an
electric circuit provided in the second semiconductor substrate
20.
[0021] As illustrated in FIGS. 1B and 1C, the semiconductor device
1 may further include a control unit which controls the
photoelectric conversion unit 11 and the signal processing unit 22.
The control unit may be provided in at least one of the first
semiconductor substrate 10 and the second semiconductor substrate
20. In the example illustrated in FIG. 1B, a control unit 12 is
provided in the first semiconductor substrate 10. In the example
illustrated in FIG. 1C, a control unit 21 is provided in the second
semiconductor substrate 20. Alternatively, a control unit for the
photoelectric conversion unit 11 may be provided in the first
semiconductor substrate 10 and another control unit for the signal
processing unit 22 may be provided in the second semiconductor
substrate 20. The control unit 12 may include a vertical driving
circuit and a power circuit. The vertical driving circuit supplies
a drive signal to a pixel circuit via a vertical scanning line. The
control unit 21 may include timing generating circuit for driving
the signal processing unit 22, a reference signal supply circuit
for supplying a reference signal to the conversion circuit, and a
horizontal scanning circuit for sequentially reading signals from
the supply circuit and the amplifier circuit.
[0022] A light control film, which is not illustrated, for
controlling light to the photoelectric conversion unit 11 may be
provided on the first semiconductor substrate 10. The light control
film may include a color filter array, a microlens array and a
light blocking layer.
[0023] In FIG. 1A, the thickness of the first semiconductor
substrate 10 is denoted by T11, the thickness of the second
semiconductor substrate 20 is denoted by T21 and the thickness of
the multilayer film 30 is denoted by T30. The thickness T11 of the
first semiconductor substrate 10 is the distance between a front
surface 103 of the first semiconductor substrate 10 and a back
surface 104 of the first semiconductor substrate 10. The thickness
T21 of the second semiconductor substrate 20 is the distance
between a front surface 203 of the second semiconductor substrate
20 and a back surface 206 of the second semiconductor substrate 20.
The thickness T30 is substantially equal to the distance between
the front surface 103 of the first semiconductor substrate 10 and
the front surface 203 of the second semiconductor substrate 20. The
distance from the second semiconductor substrate 20 and a
light-receiving surface of the first semiconductor substrate 10,
i.e., the distance between the back surface 104 of the first
semiconductor substrate 10, which is the light-receiving surface of
the first semiconductor substrate 10, and the front surface 203 of
the second semiconductor substrate 20 is denoted by T13. The
distance T13 is substantially equal to the sum of the thickness T11
and the thickness T30. The thickness T21 of the second
semiconductor substrate 20 is greater than the distance T13 between
the light-receiving surface (i.e., the back surface 104) of the
first semiconductor substrate 10, and the second semiconductor
substrate 20 (T13<T21). Desirably, the thickness T21 of the
second semiconductor substrate 20 is greater than 20 micrometers.
The thickness T21 of the second semiconductor substrate 20 is
smaller than 500 micrometers. If the second semiconductor substrate
20 is a silicon substrate, it is desirable that the thickness T21
of the second semiconductor substrate 20 is equal to or smaller
than 400 micrometers. It is desirable that the thickness T21 of the
second semiconductor substrate 20 is equal to or greater than 10
times the thickness T11 of the first semiconductor substrate 10. It
is desirable that the thickness T21 of the second semiconductor
substrate 20 is equal to or smaller than 100 times the thickness
T11 of the first semiconductor substrate 10.
[0024] An example of the semiconductor device 1 will be described
with reference to FIG. 2. FIG. 2 is a cross-sectional view of the
semiconductor device 1 along a surface including a point P and a
point Q illustrated in FIG. 1A.
[0025] The first semiconductor substrate 10 includes a photodiode
PD which includes an n-type semiconductor region 34 and a p-type
semiconductor region 35 as photoelectric conversion elements of the
photoelectric conversion unit 11. The photodiode PD is provided in
a p-type semiconductor region 32. A MOS transistor Tr1 as the
transfer transistor and a MOS transistor Tr2 as the reset
transistor are provided in the photoelectric conversion unit 11 of
the first semiconductor substrate 10. MOS transistors Tr3 and Tr4
as the semiconductor devices which constitute the control unit of
the first semiconductor substrate 10 are provided in the
photoelectric conversion unit 11. In this example, a part of the
front surface 103 of the first semiconductor substrate 10 is
constituted by the p-type semiconductor region 35 of the photodiode
PD. A part of the front surface 103 of the first semiconductor
substrate 10 forms an interface with gate dielectric films of the
MOS transistors Tr1, Tr2, Tr3 and Tr4. A device separating portion
38 is provided in the first semiconductor substrate 10. The device
separating portion 38 is provided to a deeper position of the first
semiconductor substrate 10 than the front surface 103 of the first
semiconductor substrate 10 and is made to protrude from the front
surface 103 of the first semiconductor substrate 10. Insulator
layers 43a and 43b as protective layers for protecting the front
surface 103 of the first semiconductor substrate 10 are provided on
the first semiconductor substrate 10. Insulator layers 39a, 39b,
39c, 39d and 39e are provided on the insulator layer 43b as
inter-layer insulator layers. A conductor layer 44 as a contact
plug and a plurality of conductor layers 49a, 49b and 49c as wiring
layers are provided on the first semiconductor substrate 10. The
conductor layer 44 is made of tungsten and the wiring layers 49a,
49b and 49c are made of copper. The plurality of insulator layers
43a, 43b, 39a, 39b, 39c, 39d and 39e and the plurality of conductor
layers 44, 49a, 49b and 49c constitute a first multilayer film
31.
[0026] The second semiconductor substrate 20 includes MOS
transistors Tr6, Tr7 and Tr8 as semiconductor devices which
constitute the signal processing unit. In this example, a part of
the front surface 203 of the second semiconductor substrate 20
forms an interface with gate dielectric films of the MOS
transistors Tr6, Tr7 and Tr8. A device separating portion 50 is
provided in the second semiconductor substrate 20. The device
separating portion 50 is provided to a deeper position of the
second semiconductor substrate 20 than the front surface 203 of the
second semiconductor substrate 20 and is made to protrude from the
front surface 203 of the second semiconductor substrate 20.
Insulator layers 43c and 43d as protective layers for protecting
the front surface 203 of the second semiconductor substrate 20 are
provided on the second semiconductor substrate 20. Insulator layers
49a, 49b, and 49c, 49d, 49e and 49f are provided on the insulator
layer 43d as inter-layer insulator layers. An insulator layer 59 is
provided on the inter-layer insulator layer 49f as a stress
relaxation layer. A conductor layer 54 as a contact plug and a
plurality of conductor layers 53a, 53b, 53c and 57 as wiring layers
are provided on the second semiconductor substrate 20. The
conductor layer 54 is made of tungsten, the conductor layers 49a,
49b and 49c are made of copper and the conductor layer 57 is made
of aluminum. The plurality of insulator layers 43c, 43d, 49a, 49b,
49c, 49d, 49e, 49f and 59 and the plurality of conductor layers 54,
53a, 53b, 53c and 57 constitute a second multilayer film 32. The
second multilayer film 32 covers the front surface 203 of the
second semiconductor substrate 20.
[0027] The insulator layer 39e of the first multilayer film 31 and
the insulator layer 59 of the second multilayer film 32 are bonded
via an adhesive layer 33. The first multilayer film, the second
multilayer film and the adhesive layer 33 correspond to the
multilayer film 30 illustrated in FIG. 1A. In this manner, the
first semiconductor substrate 10 and the second semiconductor
substrate 20 are bonded to each other via the multilayer film
30.
[0028] A bonding electrode 68 which penetrates the first
semiconductor substrate 10 and connects the conductor layer 49c and
the conductor layer 57 to each other is provided. The bonding
electrode 68 is made of copper or tungsten. The bonding electrode
68 is surrounded by an insulating spacer 42. The photoelectric
conversion unit 11 and the signal processing unit 22, the
photoelectric conversion unit 11 and the control unit 21, and the
control unit 12 and the signal processing unit 22 are electrically
connected by the bonding electrode 68.
[0029] A light control film 40 including antireflection layer 61, a
clad layer 62, a light blocking layer 63, a core layer 69, a
flattening layer 71, a color filter array 73 and a microlens array
74 is provided on the side of the back surface 104 of the first
semiconductor substrate 10. The refractive index of the core layer
69 is higher than the refractive index of the clad layer 62. An
optical waveguide 70 which guides light by total reflection is
formed on the core layer 69. The light control film 40 is in
contact with the p-type semiconductor region 32 which constitutes
the light-receiving surface (i.e., the back surface 104) of the
first semiconductor substrate 10. A surface 401 of the light
control film 40 on the side opposite to the first semiconductor
substrate 10 is a light input surface of the light control film. In
this example, the light input surface is constituted by the
microlens array 74.
[0030] An electrode pad 78 is arranged on a layer at the same level
as that of the conductor layer 57 and is made of aluminum. An
opening 77 which penetrates a plurality of insulator layers, the
first semiconductor substrate 10 and the light control film 40 is
provided on the electrode pad 78. A bonding wire 79 connected to
the electrode pad 78 is provided in the opening 77.
[0031] The semiconductor device 1 of this example constitutes a
back surface irradiation type photoelectric conversion device in
which a surface (i.e., the back surface 104) of the first
semiconductor substrate 10 opposite to the side on which the
transistors Tr1 to Tr4 are provided (i.e., the front surface 103)
is the light-receiving surface. The semiconductor device 1 may be a
front surface irradiation type photoelectric conversion device in
which the surface of the first semiconductor substrate 10 on which
the transistors Tr1 to Tr4 are provided (i.e., the front surface
103) is used as the light-receiving surface. In the back surface
irradiation type photoelectric conversion device, the thickness T11
of the first semiconductor substrate 10 is smaller than 10
micrometers and is typically 3 to 5 micrometers.
[0032] In FIG. 2, the thickness T11 of the first semiconductor
substrate 10, the thickness T21 of the second semiconductor
substrate 20 and the distance T30 between the first semiconductor
substrate 10 and the second semiconductor substrate 20 are
illustrated. The distance T13 between the light-receiving surface
of the first semiconductor substrate 10 and the second
semiconductor substrate 20 is the sum of the thickness T11 and the
distance T30. In this example, the thickness T12 of the first
multilayer film 31 equals to the sum of the thickness of the
insulator layers 43a, 43b, 39a, 39b, 39c, 39d and 39e. In this
example, the thickness T22 of the second multilayer film 32 equals
to the sum of the thickness of the insulator layers 43c, 43d, 49a,
49b, 49c, 49d, 49e, 49f and 59. The thickness of the adhesive layer
33 is denoted by T33. T30 is the sum of T12, T22 and T33. The sum
of the thickness T40 of the light control film 40, T13 and T21
equals to the thickness T50 of the entire semiconductor device 1
(T50=T40+T13+T21=T40+T11+T30+T21).
[0033] FIG. 2 illustrates an example in which the first multilayer
film 31 and the second multilayer film 32 are bonded using the
adhesive layer 33. However, it is also possible an insulator layer
of the first multilayer film 31 and an insulator layer of the
second multilayer film 32 are directly bonded by, for example,
plasma bonding. Alternatively, it is also possible that a conductor
layer of the first multilayer film 31 and a conductor layer of the
second multilayer film 32 are directly bonded by metal bonding.
Desirably, the conductor layer to be metal bonded is made of
copper. Thus, if direct bonding is used, the thickness T30 of the
multilayer film 30 equals to the sum of the thickness T31 of the
first multilayer film 31 and the thickness T32 of the second
multilayer film 32.
[0034] As described above, an influence of generation of heat in
the signal processing unit 22 may be reduced by setting T13<T21.
The reason thereof will be described below. The following are the
influence of generation of heat in the signal processing unit 22.
For example, heat distribution in a path to the first semiconductor
substrate 10 from the signal processing unit 22. Insulating
materials (i.e., a plurality of insulator layers) of which thermal
conductivity is generally lower than that of a semiconductor are
situated in this path. Typically, a plurality of silicon oxide
layers (thermal conductivity: 1.5 W/mK) of which thermal
conductivity is lower than that of the silicon substrate (thermal
conductivity: 150 W/mK) and a plurality of silicon nitride layers
(thermal conductivity: 30 W/mK) are situated. Thermal conductivity
mentioned here is an example.
[0035] Therefore, heat distribution in the multilayer film 30 is
that the temperature is higher in the side of the second
semiconductor substrate 20 than in the side of the first
semiconductor substrate 10. This heat distribution produces thermal
stress in the semiconductor device 1, which may cause warpage of
the semiconductor device 1 or peeling of a bonding surface (in this
example, the adhesive layer 33). Typically, the multilayer film 30
includes an electric conductor (i.e., a plurality of conductor
layers) of which thermal conductivity is generally higher than that
of a semiconductor. For the reduction of transmission of heat from
the signal processing unit 22 to the photoelectric conversion unit
11 via the electric conductor with high thermal conductivity, it is
desirable to set the density of the electric conductor in the
multilayer film 30 to be lower than the density of the insulating
material.
[0036] It has been known generation of heat in the signal
processing unit 22 mainly occurs at a position very close to the
front surface 203 of the second semiconductor substrate 20. The
cause thereof is estimated to be, for example, current leakage in
the MOS transistor. Then, on the basis of the front surface 203 of
the second semiconductor substrate 20, the relationship between T13
and T21, which are included in the thickness T50 of the
semiconductor device 1, is set to T13<T21. T13 corresponds to
the distance between the front surface 203 of the second
semiconductor substrate 20 and the light-receiving surface of the
first semiconductor substrate 10. T21 corresponds to the distance
between the front surface 203 of the second semiconductor substrate
20 and the back surface 206 of the second semiconductor substrate
20. Then, the second semiconductor substrate 20 is provided with
the mechanical strength to solve the problem described above. If
the thickness T21 of the second semiconductor substrate 20 is
greater than 50 micrometers, sufficient mechanical strength may be
obtained and, therefore, it is more desirable to set the thickness
T21 of the second semiconductor substrate 20 to be equal to or
greater than 100 micrometers.
[0037] On the other hand, if the thickness of the second
semiconductor substrate 20 is excessively large, heat in the front
surface 203 of the second semiconductor substrate 20 is not easily
radiated from the back surface 206 of the second semiconductor
substrate 20. Heat-radiating ability may be improved practically by
setting the thickness of the second semiconductor substrate 20 to
be smaller than 500 micrometers. If the second semiconductor
substrate 20 is a silicon substrate, it has been known that the
thickness T21 of the second semiconductor substrate 20 is desirably
equal to or smaller than 400 micrometers when a simulation
calculation result on the basis of thermal resistance of silicon is
considered.
[0038] Since the first semiconductor substrate 10 is situated
between the light control film 40 and the second semiconductor
substrate 20, the light control film 40 does not have a large
influence on the problem described above. However, if the stress of
the entire semiconductor device 1 is considered, it is not
desirable that the thickness T40 of the light control film 40 is
excessively large. It is desirable that T40 is not greater than the
distance T30 between the first semiconductor substrate 10 and the
second semiconductor substrate 20 (140<130). It is also
desirable that the sum of T40 and T13 is smaller than T21
(T40+T13<T21). Although the thickness T50 of the entire
semiconductor device 1 is typically smaller than 1000 micrometers,
restrictions of packaging which will be described later in the
second embodiment may be reduced by setting the thickness T50 to be
smaller than 500 micrometers. The size of the general photoelectric
conversion unit 11 in a direction perpendicular to the thickness
direction (sectional direction) of the photoelectric conversion
unit 11 (plane direction) is equal to or greater than 1.0 mm.
Especially if the size of the photoelectric conversion unit 11 in
the plane direction is equal to or greater than 5.0 mm, the
influence of heat conduction in the sectional direction is greater
than the influence of heat conduction in the plane direction.
Therefore, the present embodiment is desirable. In a typical
rectangular photoelectric conversion unit 11, the size in the plane
direction corresponds to the diagonal length.
Second Embodiment
[0039] A configuration of a photoelectric conversion device 1000
which includes a semiconductor device 1 and a package 2 will be
described with reference to FIGS. 3A-1 to 3D-2. In the present
embodiment, the package 2 includes a fixing member 3, a transparent
member 4, a support member 5 and a heat-conductive member 6.
However, this configuration is not restrictive.
[0040] In a first example illustrated in FIG. 3A-1, a signal
processing unit 22 of the semiconductor device 1 is connected to a
first terminal 81 provided in the frame-shaped fixing member 3 by a
bonding wire 79 which is connected to an electrode pad 78 of the
semiconductor device 1. The fixing member 3 constitutes the package
2. The first terminal 81 is connected to the second terminal 82 by
wiring which is not illustrated. In this example, the second
terminal 82 includes a land grid array (LGA) structure. The second
terminal 82 may include a ball grid array (BGA) structure or a pin
grid array (PGA) structure. The package 2 includes the transparent
member 4 which covers the semiconductor device 1. Although the
transparent member 4 is a plate-shaped member in this example, the
transparent member 4 may have a lens shape in order to condense the
light on the semiconductor device 1. The transparent member 4 is
supported by a support member 5 which is fixed to the fixing member
3.
[0041] Such a semiconductor device 1 of which thickness T50 is
smaller than 1000 micrometers may be incorporated in an image
pickup system as the photoelectric conversion device 1000 by using
the package 2. In the incorporation in the image pickup system, the
second terminal 82 is mounted on a mounting member which is not
illustrated. The second terminal 82 may be mounted on a front
surface of the mounting member or may be inserted in the mounting
member. Mounting on the front surface is more desirable. A printed
circuit board may be used as the mounting member. The printed
circuit board may be a rigid board, such as a glass epoxy board, a
flexible board, such as a polyimide board, or a flexible rigid
board in which a rigid board and a flexible board are combined. The
image pickup system may include a display unit for displaying image
information obtained from the photoelectric conversion device 1000,
and a recording unit for recording the image information. The
display unit may be a touch panel. The image pickup system may be a
camera, such as a still camera and a video camera. An information
terminal having camera function may also be used. The image pickup
system may include a communication unit which is connected to, for
example, the Internet and a telephone line. FIG. 3A-2 is an
enlarged view of a portion surrounded with a circle in FIG. 3A-1. A
heat-conductive member 6 is provided on the side opposite to the
first semiconductor substrate 10 via the second semiconductor
substrate 20. The heat-conductive member 6 is situated at an
opening of the frame-shaped fixing member 3 and is fixed to the
fixing member 3 at a portion which is not illustrated. Since the
heat-conductive member 6 is situated at the opening, an increase in
the thickness of the photoelectric conversion device may be
prevented.
[0042] This heat-conductive member 6 has thermal conductivity which
is higher than that of an insulator layer of a multilayer film 30
and functions as a heat-radiation member (i.e., a heat sink or a
heat spreader) for radiating heat of the semiconductor device 1.
Typically, since the most part of the insulator layer is a silicon
oxide layer, it is desirable that the heat-conductive member 6 is
made of a material having thermal conductivity equal to or greater
than 10 W/mK. If the thermal conductivity of the insulator layer is
anisotropic, the thermal conductivity of the insulator layer in the
normal direction of the second semiconductor substrate 20 should be
measured. The heat-conductive member 6 is provided in contact with
an orthogonal projection area 2061 of a photoelectric conversion
unit 11 of the first semiconductor substrate 10 on a back surface
206 which is a surface of the second semiconductor substrate 20 on
the side opposite to the first semiconductor substrate 10. If the
distance between the heat-conductive member 6 and the second
semiconductor substrate 20 is shorter than the distance T30 between
the first semiconductor substrate 10 and the second semiconductor
substrate 20, it may be considered that the heat-conductive member
6 is in contact with the second semiconductor substrate 20. For
example, the back surface 206 of the second semiconductor substrate
20 may be, in some cases, covered with an insulator film having
thermal conductivity equal to or lower than the thermal
conductivity of the insulator layer (e.g., a single layer film or a
multilayer film including a silicon oxide layer and a silicon
nitride layer). An air layer may exist between the heat-conductive
member 6 and the second semiconductor substrate 20. In these cases,
if the thickness of the insulator film or the air layer is smaller
than T30 and the distance between the heat-conductive member 6 and
the second semiconductor substrate 20 is smaller than T30, it may
be considered that the heat-conductive member 6 is in contact with
the second semiconductor substrate 20.
[0043] Desirably, the heat-conductive member 6 is a high
heat-conductive member having thermal conductivity which is higher
than the thermal conductivity of the second semiconductor substrate
20. If the second semiconductor substrate 2 is a silicon substrate,
the heat-conductive member 6 is desirably made of, for example,
aluminum (236 W/mK) and copper (398 W/mK).
[0044] Although the heat-conductive member 6 and the fixing member
3 are separated members here, these members may be integrated with
each other by using, for example, ceramic. The heat-conductive
member 6 may be a housing of the image pickup system.
[0045] As illustrated in FIG. 3A-2, the heat-conductive member 6
may be constituted by a plurality of portions. For example, the
heat-conductive member 6 may be constituted by a first
heat-conductive portion 61 having thermal conductivity which is
higher than the thermal conductivity of the insulator layer of the
multilayer film 30 and a second heat-conductive portion 62 having
thermal conductivity which is higher than the thermal conductivity
of the second semiconductor substrate 20. The thermal conductivity
of the first heat-conductive portion 61 may be lower than the
thermal conductivity of the second semiconductor substrate 20. For
example, the first heat-conductive portion 61 is provided between
the second semiconductor substrate 20 and the second
heat-conductive portion 62 and the first heat-conductive portion 61
is made to be in contact with both the second semiconductor
substrate 20 and the second heat-conductive portion 62. In this
case, the first heat-conductive portion 61 is desirably film-shaped
or sheet-shaped. Desirably, the distance TL between the first
semiconductor substrate 10 and the second heat-conductive portion
62 (i.e., the thickness of the first heat-conductive portion 61) is
smaller than the distance T30 between the first semiconductor
substrate 10 and the second semiconductor substrate 20. With this
configuration, thermal resistance between the second
heat-conductive portion 62 and the second semiconductor substrate
20 may be set smaller than the thermal resistance between the first
semiconductor substrate 10 and the second semiconductor substrate
20 (i.e., the thermal resistance of the multilayer film 30).
[0046] A second example illustrated in FIGS. 3B-1 and 3B-2 is
Modification of the first example. FIG. 3B-2 is an enlarged view of
a portion surrounded with a circle in FIG. 3B-1. The
heat-conductive member 6 (in this example, the second
heat-conductive portion 62) includes a fin structure on a front
surface 620 on the side opposite to the second semiconductor
substrate 20. Therefore, a surface area of an orthogonal projection
area 6201 from the second semiconductor substrate 20 is larger than
a surface area of the back surface 206 which is a surface on the
side of the heat-conductive member 6 of the second semiconductor
substrate 20 on the front surface 620. Since the surface area of
the orthogonal projection area 6201 is increased by the fin
structure of the heat-conductive member 6, heat may be radiated
efficiently. Although the fin structure which includes a
plate-shaped projecting portion is described herein, was mentioned
as example here, a needle point holder structure including
bar-shaped projections may be used. The second semiconductor
substrate 20, as in this example, may have variance of temperature
in a board surface when being driven. It is effective to suppose
that a high temperature portion 20H of which temperature is equal
to or greater than an average temperature and a low temperature
portion 20L of which temperature is lower than an average
temperature are provided if the high temperature portion 20H is
situated in the orthogonal projection area from the photoelectric
conversion unit 11.
[0047] FIG. 3B-1 illustrates an example in which the image pickup
system in which the photoelectric conversion device is incorporated
includes a fan 7 as a cooling unit. The fan 7 may carry out forced
cooling of the second semiconductor substrate 20 by carrying out
forced cooling of the heat-conductive member 6 with air cooling.
Water-cooling may be used for the forced cooling.
[0048] In a third example illustrated in FIG. 3C-1, between the
high temperature portion 20H and the low temperature portion 20L of
the second semiconductor substrate 20, the low temperature portion
20L is provided in the orthogonal projection area of the
photoelectric conversion unit 11. FIG. 3C-2 is an enlarged view of
a portion surrounded with a circle in FIG. 3C-1. At the design
phase of a sensor device, site(s) and portion(s) at which heat is
easily generated may be largely specified by simulation when the
layout and the floor plan of the circuit block are determined.
Generally, a horizontal scanning circuit and a counter circuit
easily become the high temperature portion 20H. A surface area of
the back surface 206 is large in an area 620H corresponding to the
high temperature portion 20H and the surface area of the back
surface 206 is smaller in an area 620L corresponding to the low
temperature portion 20L than in the area 620H. With this
configuration, variance of temperature in the second semiconductor
substrate 20 is averaged.
[0049] In a fourth example illustrated in FIG. 3D-1, a
heat-conductive member 63 and a heat-conductive member 64 are
connected to the first semiconductor substrate 10. This example is
desirable in a case in which the first semiconductor substrate 10
includes a control unit 12 and the control unit 12 becomes the high
temperature portion 10H of which temperature is higher than that of
the photoelectric conversion unit 11. FIG. 3D-2 is an enlarged view
of a portion surrounded with a circle in FIG. 3D-1. If the
heat-conductive member 64 is disposed on the side of the first
semiconductor substrate 10, it is important not to disturb
incidence of light. If a lens is disposed in front of the
semiconductor device 1, it is possible to know the maximum angle of
incident light in accordance with the eye relief between the lens
and the semiconductor device 1. Therefore, it is necessary to
design the radiation mechanism in consideration of the angle.
[0050] FIG. 4A illustrates an example in which a recess 216 is
formed in the second semiconductor substrate 20, the first
heat-conductive portion 61 is disposed inside the recess 216 and
the second heat-conductive portion 62 which is in contact with the
first heat-conductive portion 61 is further provided.
[0051] FIG. 4B illustrates an example in which no heat-conductive
member 6 is provided and a plurality of recesses 216 are provided
on the back surface 206 of the second semiconductor substrate 20.
In this manner, radiation efficiency from the second semiconductor
substrate 20 may be improved by setting the surface area of the
back surface 206 of the second semiconductor substrate 20 greater
than the surface area of a front surface 203 of the second
semiconductor substrate 2. Desirably, the plurality of recesses 216
have a fin structure or a needle point holder structure.
[0052] In such a configuration in which no heat-conductive member 6
is provided, forced cooling of the second semiconductor substrate
20 may be carried out by air cooling.
Third Embodiment
[0053] A method of manufacturing a photoelectric conversion device
will be described with reference to FIGS. 5A to 5J.
Process A
[0054] Photoelectric conversion element groups (not illustrated)
are formed in a plurality of areas on a front surface 1030 (i.e.,
front face) of the first semiconductor wafer 100 using a publicly
known semiconductor device formation technique, such as ion
implantation. Each of the photoelectric conversion element groups
is constituted by a plurality of arranged photoelectric conversion
elements (see FIG. 5A). The shape of a single crystal silicon wafer
which is a typical semiconductor wafer is defined by the
Semiconductor Equipment and Materials International (SEMI)
standard. Hereinbelow, examples of dimensions of the silicon wafer
practically used for the manufacture of the photoelectric
conversion device among those defined by the SEMI standard will be
described. In a wafer of which diameter is 150.000 (.+-.0.20) mm,
the thickness of the wafer center (center point) is 675 (.+-.20)
micrometers. In a wafer of which diameter is 200.000 (.+-.0.20) mm,
the thickness of the wafer center (center point) is 725 (.+-.20)
micrometers. In a wafer of which diameter is 300.000 (.+-.0.20) mm,
the thickness of the wafer center (center point) is 775 (.+-.20)
micrometers. The values with .+-. in the parenthesis are common
differences. Either of these thickness values may be used as the
thickness T10 of the first semiconductor wafer 100.
Process B
[0055] After the process A, a first multilayer film 110 which
includes a plurality of insulator layers and a plurality of
conductor layers is formed on the front surface 1030 of the first
semiconductor wafer 100 using a publicly known multilayer
interconnection technique. The thickness of the first multilayer
film 110 is T12. T12 is typically smaller than T10 (T12<T10).
All the layers of the plurality of conductor layers may be made of
the same material or made of different materials. For example,
aluminum layers and copper layers may be used in a mixed manner. In
this manner, the first semiconductor wafer 100 and a first member
111 which includes the first multilayer film 110 stacked on the
first semiconductor wafer 100 are prepared (see FIG. 5B). The
thickness of the first member 111 is the sum of T10 and T12.
Process C
[0056] Semiconductor device groups (not illustrated) are formed in
a plurality of areas on a front surface 2030 (i.e., front face) of
the second semiconductor wafer 200 using a publicly known
semiconductor device formation technique, such as ion implantation.
Each of the plurality of semiconductor device groups is constituted
by arranged semiconductor devices (see FIG. 5C). The thickness T20
of the second semiconductor wafer 200 may be the thickness defined
by the SEMI standard like the thickness T10 of the first
semiconductor wafer 100. Desirably, the first semiconductor wafer
100 and the second semiconductor wafer 200 are the same in
diameter.
Process D
[0057] After the process C, a second multilayer film 210 which
includes a plurality of insulator layers and a plurality of
conductor layers is formed on the front surface 2030 of the second
semiconductor wafer 200 using a publicly known multilayer
interconnection technique. The thickness of the second multilayer
film 210 is T22. T22 is typically smaller than T20 (T22<T20).
All the layers of the plurality of conductor layers may be made of
the same material or made of different materials. For example,
aluminum layers and copper layers may be used in a mixed manner. In
this manner, the second semiconductor wafer 200 and a second member
222 which includes the second multilayer film 210 stacked on the
second semiconductor wafer 200 are prepared (see FIG. 5D). The
thickness of the first member 111 is the sum of T10 and T12. The
order of the process B and the process D is not particularly
limited: the process B and the process D may be performed in
parallel.
Process E
[0058] Preparation for a bonding process described in the next
process E is made in at least one of the first member 111 and the
second member 222. In this example, an adhesive is applied to at
least one of a front surface of the first multilayer film 110 or a
front surface of the second multilayer film 220 which are bonding
surfaces of the first member 111 and the second member 222. In
plasma bonding is carried out, at least one of an insulator layer
of the surface of the first multilayer film 110 and an insulator
layer of the surface of the second multilayer film 220 is subject
to plasma processing. The first member 111 and the second member
222 are stacked to each other with the first multilayer film 110
and the second multilayer film 120 being disposed between the first
semiconductor wafer 100 and the second semiconductor wafer 200 (see
FIG. 5E). At this time, the first member 111 and the second member
222 are aligned with each other using alignment marks which are
previously formed at both the first member 111 and the second
member 222.
Process F
[0059] After the process E, the first member 111 and the second
member 222 are bonded to each other while keeping the state in
which the first multilayer film 110 and the second multilayer film
120 are disposed between the first semiconductor wafer 100 and the
second semiconductor wafer 200. In this example, as described in
the process E, since the adhesive which is not illustrated is used
for bonding the first member 111 and the second member 222, the
first member 111 and the second member 222 are bonded by a
solidification process of the adhesive. The solidification process
may be a photo-curing process, a heat-curing process and a drying
process. Pressure may be applied to the first member 111 and to the
second member 222 if necessary. By this process F which is the
bonding process, a composite member 330 in which the first member
111 and the second member 222 are bonded to each other is
manufactured. The sum of the thickness T12 of the first multilayer
film 110, the thickness of the adhesive layer and the thickness T22
of the second multilayer film 210 equals to the distance T30
between the first semiconductor wafer 100 and the second
semiconductor wafer 200.
Process G
[0060] After the process F, the first semiconductor wafer 100 of
the composite member 330 is thinned from the side of a back surface
1040 of the first semiconductor wafer 100. Then, a composite member
331 of which first semiconductor wafer is thinned than that of the
composite member 330 is obtained. A method of thinning the first
semiconductor wafer 100 may be, for example, chemical machinery
polishing (CMP), mechanical polishing (MP), wet etching and dry
etching. These methods may be used in combination. By this process
F which is a first thinning process, the first semiconductor wafer
100 which has had the thickness T10 until the process F becomes a
first semiconductor wafer 101 which has the thickness T11
(T11<T10). The first semiconductor wafer 100 which includes the
back surface 1040 becomes the first semiconductor wafer 101 which
includes a back surface 1041. The distance T13 between a front
surface of the second semiconductor wafer 200 (i.e., a front
surface 2030) and a front surface of the first semiconductor wafer
110 (i.e., a front surface 1030) equals to the sum of the distance
T30 and the thickness T11. The thickness T11 of the first
semiconductor wafer 101 is equal to or smaller than 10 micrometers
and typically is 3 to 5 micrometers.
Process H
[0061] After the process G, bonding electrode(s) (not illustrated)
for connecting the conductor layer of the first multilayer film 110
and the conductor layer of the second multilayer film 210 are
formed if necessary. The bonding electrodes may be provided by
penetrating the first semiconductor wafer 101 or, alternatively, by
previously removing a part of the first semiconductor wafer 101,
disposing the insulator layer at the portion from which the first
semiconductor wafer 101 is removed, and penetrating the insulator
layer. Details of the formation method of the bonding electrodes
are provided in Japanese Patent Laid-Open No. 2011-96851, Japanese
Patent Laid-Open No. 2011-151375 and Japanese Patent Laid-Open No.
2011-204915.
[0062] A light control film 400 which includes at least one of a
color filter array, a microlens array and a light blocking layer is
formed on the back surface 1041 of the first semiconductor wafer
101. Then, a composite member 332 of which thickness is greater
than that of the composite member 331 by the thickness of the light
control film 400 is obtained. The color filter array may include a
light transmission unit which transmits red light, green light and
blue light which are the primary lights and, in addition to that, a
light transmission unit which transmits white light or
complementary color light of the primary lights. The microlens
array may be formed by reflowing or by etchback. However, gradual
exposure is desirable from the viewpoint of reducing the damage to
the first semiconductor wafer 101. The thickness T40 of the light
control film 400 is not particularly limited. Desirably, however,
T40 is not greater than the distance T30 between the first
semiconductor wafer 101 and the second semiconductor wafer 200
(T40<T30). Although the process of forming the light control
film 400 is desirably performed after the process of forming the
bonding electrode(s), these processes may be performed in reverse
order. A surface 4010 of the light control film 400 becomes a light
input surface.
Process I
[0063] After the process H, the second semiconductor wafer 200 of
the composite member 332 is thinned from a back surface 2060 side
of the second semiconductor wafer 200. Then, a composite member 333
of which second semiconductor wafer 200 is thinner than that of the
composite member 332 is obtained. A method of thinning the second
semiconductor wafer 200 may be, for example, chemical machinery
polishing (CMP), mechanical polishing (MP), wet etching and dry
etching. These methods may be used in combination. By this process
I which is a second thinning process, the second semiconductor
wafer 200 which has had the thickness T20 until the process H
becomes a second semiconductor wafer 201 which has the thickness
T21 (T21<T20). Desirably, the thickness T21 is smaller than 500
micrometers. The second semiconductor wafer 200 which includes the
back surface 2060 becomes the second semiconductor wafer 201 which
includes a back surface 2061. Typically, the thickness T21 of the
second semiconductor wafer 201 after the thinning process is equal
to or smaller than half the thickness T20 of the second
semiconductor wafer 200 before the thinning process. If the second
semiconductor wafer 200 is a silicon wafer, the thickness T21 of
the second semiconductor wafer 201 is desirably equal to or smaller
than 400 micrometers. It is important to set the thickness T21 to
be greater than the distance T13 between the back surface 1041 of
the first semiconductor wafer 101 and the front surface 2030 of the
second semiconductor wafer 200 (i.e., the second semiconductor
wafer 201). If the damage to the semiconductor device group formed
in the second semiconductor wafer is considered in accordance with
the result of the X-ray topography analysis, the thickness T21 of
the second semiconductor wafer 201 is desirably equal to or greater
than 20 micrometers, and is more desirably equal to or greater than
50 micrometers.
[0064] If the thickness of the second semiconductor wafer is
relatively thick (for example, equal to or greater than 300
micrometers), the thinning process is desirably performed by
mechanical polishing in consideration of the thinning speed. On the
other hand, if the thickness of the second semiconductor wafer is
relatively thin (for example, less than 300 micrometers), the
thinning process is desirably performed by chemical machinery
grinding in consideration of the damage to the semiconductor device
group by the thinning. For example, if a silicon wafer which is 300
mm in diameter is used as a second semiconductor wafer 200,
mechanical polishing is used in a stage of thinning 775 micrometers
to 300 micrometers and chemical machinery grinding is used in a
stage of thinning from 300 micrometers to 50 micrometers. Further
thinning may be performed by wet etching.
Process J
[0065] After the process I, the composite member 333 is divided
into a plurality of semiconductor devices 1 each including a
photoelectric conversion element group and a semiconductor device
group. The dividing process of this dividing process may be
performed by blade dicing using a dicing blade (a dicing saw).
Laser dicing using a laser beam may also be used. Although the
dividing process is desirably performed from the first
semiconductor wafer 101 side of the composite member 333, it is
also possible to be performed from the second semiconductor wafer
201 side. By performing the dividing process while the relationship
of 121>T13 is satisfied, the second semiconductor wafer 201
substantially supports the rigidity of the composite member 333,
whereby the dividing process may be performed suitably. It is also
possible to support the rigidity by increasing the thickness of the
first semiconductor wafer 101. However, since a plurality of
photoelectric conversion element groups are disposed on the first
semiconductor wafer 101, suitable thickness T11 of the first
semiconductor wafer 101 is determined by the photoelectric
conversion performance. Therefore, there is a limit to the increase
in thickness for the enhancement of rigidity. On the other hand,
since the second semiconductor wafer 201 has less such limitation,
it is desirable to increase the thickness of the second
semiconductor wafer 201. It is also possible to provide a support
member for supporting the composite member 333 at the time of the
dividing process if necessary. If the thickness T50 of the
composite member 333 is equal to or greater than 50 micrometers,
the dividing process may be performed with a possibility that
chipping or peeling of the bonding surface is caused being reduced
as much as possible even if the blade dicing is used. If thickness
T50 of the composite member 333 is less than 50 micrometers, a
possibility that chipping or peeling of the bonding surface is
caused is reduced when the laser dicing is used.
Process K
[0066] After the process J, packaging of each of the plurality of
semiconductor devices 1 is carried out. Heat-conductive die bond
paste is applied to at least one of the heat sink as the second
heat-conductive portion 62, and the back surface 206 of the second
semiconductor substrate 20 of the semiconductor device 1
illustrated in FIG. 3A-1, the dye bond paste is solidified, and
then a heat-conductive film as the first heat-conductive portion 61
is formed. Then, the back surface 206 is fixed to the heat sink and
thermal connection between the heat sink and the semiconductor
device 1 is obtained in addition to mechanical connection.
[0067] Formation of the recesses 216 illustrated in FIGS. 4A and 4B
may be performed after the thinning process of the second
semiconductor wafer 200. For example, after the second thinning
process, a silicon oxide film is formed by, for example, CVD on the
back surface 2061 of the second semiconductor wafer 201. Then, the
silicon oxide film is patterned at a desired position using a
resist pattern by dry etching. Next, the silicon is etched to
desired depth by wet etching using the patterned silicon oxide film
as a hard mask using, for example, an alkaline etchant, such as
TMAH and KOH. Alternatively, the recesses 216 may be formed by dry
etching in the silicon substrate by repeating etching using gas of
which principal component is SF.sub.6 and etching using gas of
which principal component is C.sub.4F.sub.6.
[0068] While the disclosure has been described with reference to
exemplary embodiments, it is to be understood that the disclosure
is not limited to the disclosed exemplary embodiments. The scope of
the following claims is to be accorded the broadest interpretation
so as to encompass all such modifications and equivalent structures
and functions.
* * * * *